Tag der Offen Tür

KIDS Künstliche Intelligenz & Data Science

Project Cars

Goals

• Playing with real data ,make sense out it, find correlations between values.
• Using specicif features of the libraries studied in class to strengthen the knowledge

Dataset

• https://www.kaggle.com/atechnohazard/battery-and-heating-data-in-real-driving-cycles?select=TripA01.csv

Problems

• The shear size of the data set. Creating a full pairplot of all 48 columns takes more than 15 minutes, even with optimized rendering
• Understanding the Dataset the huge amount of data was enourmous to wrap our heads arround

---

#All libraries that are uses somewhere
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from IPython.display import Image
import time
import math
from IPython.display import display, HTML
import numpy as np
import ipywidgets as widgets
from sklearn.linear_model import LinearRegression

trip = pd.read_csv('./TripB01.csv') # Get First Trip
trip = trip[trip['Time [s]'] % 1 == 0] # Get only every 10th row (1s,2s,3s,4s,5s....)
trip 

	Time [s]	Velocity [km/h]	Elevation [m]	Throttle [%]	Motor Torque [Nm]	Acceleration [m/s^2]	Regenerative Braking Signal	Battery Voltage [V]	Battery Current [A]	Battery Temp	...	Temp Footweel Driver	Temp Footweel Co-Driver	Temp Feetvent Co-Driver	Temp Feetvent Driver	Temp Head Co-Driver	Temp Head Driver	Temp Vent right	Temp Vent central right	Temp Vent central left	Temp Vent right
0	0.0	0.00000	449.00000	0.0	0.00000	-0.32930	0.0	388.80000	-19.06128	7.0	...	9.30159	8.42878	7.73053	8.07965	12.53101	13.49111	10.26169	9.65072	9.65072	10.34897
10	1.0	0.00000	449.00000	0.0	0.00000	-0.42098	0.0	388.80000	-18.21941	7.0	...	9.30159	8.49573	7.99832	8.37524	12.53101	13.49111	10.26169	9.65072	9.65072	10.34897
20	2.0	0.00000	449.00000	0.0	0.00000	-0.35800	0.0	388.90000	-17.74028	7.0	...	9.30159	8.71690	8.41439	8.79826	12.53101	13.34336	10.26169	9.71767	9.65072	10.34897
30	3.0	0.00000	449.00000	0.0	0.00000	-0.34662	0.0	388.95974	-17.27922	7.0	...	9.30159	8.91179	8.91773	9.53025	12.53101	13.31655	10.26169	9.67106	9.58376	10.27508
40	4.0	0.00000	449.00000	0.0	0.00000	-0.35230	0.0	388.95976	-17.07927	7.0	...	9.37547	9.22024	9.64223	10.38966	12.45714	13.31655	10.26169	9.65072	9.56344	10.26169
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
32470	3247.0	0.57881	476.00966	0.0	4.35177	0.21247	0.0	373.54784	-3.63922	15.0	...	24.13947	24.86533	38.10452	36.27161	22.51915	22.15101	29.78516	27.63073	29.11452	28.22271
32480	3248.0	0.09375	476.00000	0.0	0.00000	-0.28795	0.0	373.60000	-3.58716	15.0	...	24.13947	24.83772	38.10452	36.27161	22.48112	22.13199	29.81277	27.69040	29.11452	28.17342
32490	3249.0	0.00000	476.00000	0.0	0.00000	-0.34132	0.0	373.60000	-3.86484	15.0	...	24.07120	24.83772	38.10452	36.27161	22.41285	22.13199	29.81277	27.71801	29.05483	28.22270
32500	3250.0	0.00000	476.00000	0.0	0.00000	-0.36000	0.0	373.60000	-3.20841	15.0	...	24.05219	24.83772	38.10452	36.27161	22.39384	22.13199	29.75311	27.71801	29.08692	28.17342
32510	3251.0	0.00000	476.00000	0.0	0.00000	-0.34236	0.0	373.60000	-3.23119	15.0	...	23.98393	24.89738	38.04486	36.20333	22.32558	22.13199	29.72549	27.71801	29.11452	28.22270

3252 rows × 48 columns

def getTrip(n,m,x,y):
    return trip.iloc[n:m,x:y]

def smallTrip(n):
    return getTrip(0+n,100+n,0,48)
    

trip.isnull().sum()

Time [s]                          0
Velocity [km/h]                   0
Elevation [m]                     0
Throttle [%]                      0
Motor Torque [Nm]                 0
 Acceleration [m/s^2]             0
Regenerative Braking Signal       0
Battery Voltage [V]               0
Battery Current [A]               0
Battery Temp                      0
max. Battery Temp                 0
SoC                               0
displayed SoC [%]                 0
min. SoC [%]                      0
max. SoC [%)                      0
Heating Power CAN [kW]            0
Heating Power LIN [W]             0
Requested Heating Power [W]       0
AirCon Power [kW]                 0
Heater Signal                     0
Heater Voltage [V]                0
Heater Current [A]                0
Ambient Temp                      0
Ambient Temp Sensor               0
Coolant Temp Heatercore           0
Requested Coolant Temp            0
Coolant Temp Inlet                0
Coolant Volume Flow +500 [l/h]    0
Heat Exchanger Temp               0
Cabin Temp Sensor [C]             0
Temp Coolant Heater Inlet         0
Temp Coolant Heater Outlet        0
Temp Heat Exchanger Outlet        0
Temp Defrost lateral left         0
Temp Defrost lateral right        0
Temp Defrost central              0
Temp Defrost central left         0
Temp Defrost central right        0
Temp Footweel Driver              0
Temp Footweel Co-Driver           0
Temp Feetvent Co-Driver           0
Temp Feetvent Driver              0
Temp Head Co-Driver               0
Temp Head Driver                  0
Temp Vent right                   0
Temp Vent central right           0
Temp Vent central left            0
Temp Vent right                   0
dtype: int64

# Function to create an image
# @param1 string -> name of the image
# @param2 sns_plot -> the plot to render
# @return The renderd image


def makeImage(name,plot):
    path='./savedImages/'+name+".png"
    
    # save the image and stop the rendering time
    start = time.time() 
    saveImage(path,plot)
    end = time.time()
    
    print("Render Time :"+str(math.ceil((end - start)*100)/100)+"s")
    print("FilePath: "+path)
    
    return Image(path) # The rendered image

def saveImage(path,plot):
    plot.savefig(path)
    plt.clf()

# Some blocks will be so big that they will have a scrollbar, I dont like the scrollbar
def clear():
    js = "<script>$('.output_scroll').removeClass('output_scroll')</script>"
    display(HTML(js))

# make the plots look better
sns.set_style("darkgrid")

plot = sns.pairplot(getTrip(0,2000,0,6), height=2.0)
makeImage("MassPlot",plot) # Show pairplot as image

Render Time :2.26s
FilePath: ./savedImages/MassPlot.png

<Figure size 864x864 with 0 Axes>

fig, ax = plt.subplots(figsize=(40,40))         # Sample figsize in inches
sns.heatmap(trip.corr(), annot=True, linewidths=.5, ax=ax)

<AxesSubplot:>

---

Conclusion of the Plots

We see that we don't see a lot of clear structure/correlation within our dataset. The next logical thing is to to think about correlations and test them out ourself.

• Time - Battery Heat 1:1
• Battery Voltage - Battery Current -1:1
• Heat of all wheels 1:1

# Time - Battery Heat

plot = sns.pairplot(trip[['Time [s]', 'Battery Temp']], height=4.0)
makeImage("TimeBatteryHeatPlot",plot) # Show pairplot as image

# Yes there is a correlation between Time and  Battery heat

Render Time :0.35s
FilePath: ./savedImages/TimeBatteryHeatPlot.png

<Figure size 576x576 with 0 Axes>

#Battery Voltage - Battery Current

bv="Battery Voltage [V]"
bc="Battery Current [A]"

plot = sns.pairplot(trip[[bv,bc]], height=4.0)
makeImage("BatteryVoltageCurrentPlot",plot) # Show pairplot as image

# Yes there is a correlation between Time and  Battery heat

Render Time :0.46s
FilePath: ./savedImages/BatteryVoltageCurrentPlot.png

<Figure size 576x576 with 0 Axes>

bv="Time [s]"
bc="Battery Current [A]"

def overview(plot,xVal,yVal):
    sns.pairplot(data=plot)
    plt.figure()
    sns.stripplot(data=plot, x=xVal, y=yVal)
    plt.figure()
    sns.jointplot(data=plot, x=xVal, y=yVal)
    plt.figure(figsize=(15,5))
    line_plot = sns.lineplot(data=plot, x=xVal, y=yVal,dashes=False)
    xticks = line_plot.get_xticks()
    line_plot.set_xticks(xticks[::40]);
    plt.figure(figsize=(15,5))
    sns.lineplot(data=plot, dashes=False)
    plt.figure()
    sns.lineplot(data=plot, dashes=False)

overview(smallTrip(0)[[bc,bv]],bv,bc)

<Figure size 432x288 with 0 Axes>

bv="Time [s]"
bc="Battery Current [A]"

overview(smallTrip(0)[[bc,bv]],bv,bc)

<Figure size 432x288 with 0 Axes>

---

Using even More data

From now on, its even more important to think before run any code
the dataset contains now 193,350 rows but we reduced it down to 1 10th, but it still stores more than 9.4 million values (19335 rows × 49 columns)

How is the new data structured ?

First of all the headers do not include the units anymore for example from now on the time coloumn is called 'Time' and not 'Time [s]' anmore

To differntiate the specific trips from each other they all of now a tripId (the last coloumn), we can NOT use it as an Id since there are 19 thousand rows and just 8 trips.

Each row has an own id created by the dataset

frames = []
for x in range(2,10):
    newTrip = pd.read_csv('./data/TripB0'+str(x)+'.csv') #read csv
    newTrip = newTrip[newTrip['Time'] % 1 == 0] # Get only every 10th row (1s,2s,3s,4s,5s....)
    newTrip["tripId"]=(x-1)
    frames.append(newTrip)
    
bigTrip = pd.concat(frames) # join all dataframes together
bigTrip

	Time	Velocity	Elevation	Throttle	Motor Torque	Longitudinal Acceleration	Regenerative Braking Signal	Battery Voltage	Battery Current	Battery Temperature	...	Temperature Footweel Co-Driver	Temperature Feetvent Co-Driver	Temperature Feetvent Driver	Temperature Head Co-Driver	Temperature Head Driver	Temperature Vent right	Temperature Vent central right	Temperature Vent central left	Temperature Vent right.1	tripId
0	0.0	0.0	437.0	0.0	0.0	-0.34	0.0	389.9	-1.60	13.0	...	9.91	12.44	10.17	11.31	11.40	12.01	11.13	11.48	10.87	1
10	1.0	0.0	437.0	0.0	0.0	-0.30	0.0	389.9	-1.54	13.0	...	9.88	12.44	10.21	11.27	11.52	12.01	11.17	11.48	10.83	1
20	2.0	0.0	437.0	0.0	0.0	-0.34	0.0	389.9	-1.54	13.0	...	9.83	12.48	10.30	11.22	11.74	11.97	11.22	11.48	10.75	1
30	3.0	0.0	437.0	0.0	0.0	-0.31	0.0	389.9	-1.44	13.0	...	9.79	12.56	10.31	11.22	11.79	11.92	11.19	11.48	10.66	1
40	4.0	0.0	437.0	0.0	0.0	-0.33	0.0	389.9	-1.46	13.0	...	9.70	12.65	10.22	11.26	11.71	11.92	11.13	11.48	10.65	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
56060	5606.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	25.89	39.33	37.93	21.70	22.66	32.87	32.43	33.04	30.95	8
56070	5607.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	25.89	39.40	38.02	21.70	22.74	32.87	32.43	32.96	30.95	8
56080	5608.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	25.42	39.10	37.67	21.78	22.66	32.56	32.35	32.80	30.95	8
56090	5609.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	24.58	36.57	35.35	21.70	22.48	32.28	32.34	32.62	30.86	8
56100	5610.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	24.49	36.27	32.99	21.70	22.40	32.26	32.34	32.61	30.86	8

19335 rows × 49 columns

bigTrip.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 19335 entries, 0 to 56100
Data columns (total 49 columns):
 #   Column                             Non-Null Count  Dtype  
---  ------                             --------------  -----  
 0   Time                               19335 non-null  float64
 1   Velocity                           19335 non-null  float64
 2   Elevation                          19335 non-null  float64
 3   Throttle                           19335 non-null  float64
 4   Motor Torque                       19335 non-null  float64
 5   Longitudinal Acceleration          19335 non-null  float64
 6   Regenerative Braking Signal        19335 non-null  float64
 7   Battery Voltage                    19335 non-null  float64
 8   Battery Current                    19335 non-null  float64
 9   Battery Temperature                19335 non-null  float64
 10  max. Battery Temperature           19335 non-null  float64
 11  SoC                                16863 non-null  float64
 12  displayed SoC                      16863 non-null  float64
 13  min. SoC                           16863 non-null  float64
 14  max. SoC                           16863 non-null  float64
 15  Heating Power CAN                  19335 non-null  float64
 16  Heating Power LIN                  19335 non-null  float64
 17  Requested Heating Power            19335 non-null  float64
 18  AirCon Power                       19335 non-null  float64
 19  Heater Signal                      19335 non-null  int64  
 20  Heater Voltage                     19335 non-null  float64
 21  Heater Current                     19335 non-null  float64
 22  Ambient Temperature                19335 non-null  float64
 23  Ambient Temperature Sensor         18916 non-null  float64
 24  Coolant Temperature Heatercore     19335 non-null  float64
 25  Requested Coolant Temperature      19335 non-null  int64  
 26  Coolant Temperature Inlet          19335 non-null  float64
 27  Coolant Volume Flow +500           16545 non-null  float64
 28  Heat Exchanger Temperature         19335 non-null  float64
 29  Cabin Temperature Sensor           19335 non-null  float64
 30  Temperature Coolant Heater Inlet   18916 non-null  float64
 31  Temperature Coolant Heater Outlet  18916 non-null  float64
 32  Temperature Heat Exchanger Outlet  18916 non-null  float64
 33  Temperature Defrost lateral left   18916 non-null  float64
 34  Temperature Defrost lateral right  18916 non-null  float64
 35  Temperature Defrost central        18916 non-null  float64
 36  Temperature Defrost central left   18916 non-null  float64
 37  Temperature Defrost central right  18916 non-null  float64
 38  Temperature Footweel Driver        18916 non-null  float64
 39  Temperature Footweel Co-Driver     18916 non-null  float64
 40  Temperature Feetvent Co-Driver     18916 non-null  float64
 41  Temperature Feetvent Driver        18916 non-null  float64
 42  Temperature Head Co-Driver         18916 non-null  float64
 43  Temperature Head Driver            18916 non-null  float64
 44  Temperature Vent right             18916 non-null  float64
 45  Temperature Vent central right     18916 non-null  float64
 46  Temperature Vent central left      18916 non-null  float64
 47  Temperature Vent right.1           18916 non-null  float64
 48  tripId                             19335 non-null  int64  
dtypes: float64(46), int64(3)
memory usage: 7.4 MB

Hier kann man nun sehen dass wir auch einige (Tausende) Nullfelder haben dies wird sich auf die Qualität der nächsten Plots auswirken jedoch gehe ich nicht davon aus das die Qualität sehr stark darunter leiden wird da es sich um ungefähr 1-2% null felder handelt

---

Time - Acceleration Correlation

 
sns.lmplot(x='Time', y='Longitudinal Acceleration', data=bigTrip[0:60])

<seaborn.axisgrid.FacetGrid at 0x213ea1aed30>

Hier können wir sehen, dass wie davon schon auszugehen war die Beschleunigung innerhalb der ersten Minute nach Start eines Trips immer größer wird, wenn wir diesen Graphen nun von 60 Sekunden auf eine Stunde Zeichnen lassen würden ,wäre unser Ergebnis vermutlich um einiges verstreuter, weil während eines Trips auch gebremst werden muss (Stau, Ampel, Personen auf der Fahrbahn)

Und genau dass schauen wir uns jetzt an

sns.lmplot(x='Time', y='Longitudinal Acceleration', data=bigTrip[0:1000])
plt.figure()
sns.lmplot(x='Time', y='Longitudinal Acceleration', data=bigTrip[1570:1600])

<seaborn.axisgrid.FacetGrid at 0x213e9f590d0>

<Figure size 432x288 with 0 Axes>

Wie schon erwähnt ist dieser Graph sehr verstreut, trotzdem kann man auch in diesem Graphen erkennen das die Beschleunigung zu Begin der Fahrt eher größer wird als zum Ende hin

Der zweite Graph zeit die letzen 30 Sekunden der Trips die Beschleunigung nimmt stark ab.

sns.lmplot(x='Time', y='Heating Power LIN', data=bigTrip[0:1600])

<seaborn.axisgrid.FacetGrid at 0x213e9f30550>

Beobachtung

In diesen Graph kann man sehr gut erkennen, dass die benötigte Energie zum heizen in den ersten 160 Sekungen immer höher wird und dann ab ungefähr 200 Sekunden starb abnimmt und fortan gleich bleibt.

Mögliche Erklräung

Dies liegt vermutlich daran ,dass die Heizung eines Autos normalerweise über die Wärme des Motors reguliert wird, zu begin der Fahrt ist der Motor normalerweise nicht recht warm/heiß weswegen eine zusätzliche Heizung sprich die zuheizung eingschaltet werden muss, diese verbraucht enorm viel energie was sich im Graphen auch erkenntlich zeigt.

---

Wheel Temperature during a Trip

sns.lmplot(x='Time', y='Temperature Footweel Driver', data=bigTrip[0:1600])

<seaborn.axisgrid.FacetGrid at 0x213ea09a370>

Beobachtung

Die Reifen Temperatur steigt in den ersten 3 Minuten stark an. Danach schwankt die Temperatur mit einer trotzdem steingenden Tendez.

Da die durchscnittliche Start Reifentemperatur bei ~10.2 Grad Celsisus liegt, können wir eventuell davon ausgehen das die Messungen eher im Winter durchgeführt wurden.

Mögliche Erklärung

Durch Reibung ensteht Wärme aufgrund von Thermodynamik. Die Schwankungen können eventuell auf Rote Ampeln, andere Straßen verhältnisse oder auf Stau's zurückgeführt werden

---

Coolant Temperature during a Trip

sns.set_theme(style="whitegrid")

# Initialize the matplotlib figure
f, ax = plt.subplots(figsize=(12, 6))

# Only calculate the sample once
data=bigTrip[bigTrip['Time'] % 15 == 0 ][0:15]

# Plot the total crashes
sns.set_color_codes("pastel")
sns.barplot(x="Time", y="Coolant Temperature Inlet", data=data,
            label="Total", color="b")

plt.figure()

sns.set_color_codes("muted")
sns.barplot(x="Time", y="Temperature Coolant Heater Outlet", data=data,
             label="Alcohol-involved", color="b")

<AxesSubplot:xlabel='Time', ylabel='Temperature Coolant Heater Outlet'>

Datensatz

Für diese Graphen wurden nur 15 Datensätze genommen um einen Möglichst "schönen" Graphen zu erhalten. Um Trotzdem eine großen Datenbereich abzudecken wurden Datensätze mit "großen" Zeitlichen Abstand genommen, sprich ein Datensatz alle 15 Sekunden was sich in diesen Fall auf die ersten 3 1/2 Minunten nach dem Start eines Trips.

Das nutzen von Datensätzen von späteren Zeitpunkte würde vermutlich weniger Sinn ergeben, weil das Kühlmittel die maximal Temperatur* erreicht und dann nur noch schwache veränderungen zeigt.

*die Temperatur die das Kühlmittel Maximal aufgrund der Motorwärme m in einem Abstand zum Motor n erreichen kann

Beobachtung

Die Kühlmittelflüssigkeit wird im Laufe der Zeit immer Wärmer.

Mögliche Erklärung

Eine Erklärung dafür, dass die Eingangstemperatur niedrieger ist als die Ausgangtemperatur. Zu beginn muss man erwähnen dass die Flüssigkeit zu jeder Zeit im system ist, sprich es handelt sich hier nicht um das gleiche Prinzip wie bei einem Wärmetauscher.

Ich denke, dass die nähe zum Motor die bei der Seite des Ausgangs näher ist spielt hier eine wessentliche Rolle. Der Motor strahlt innerhalb des gesamten Motorraumes wärme aus und diese flacht mit der distanz ab.

Die obigen Sätze als Bild zusammengefasst :

Data Training Regression

bigTrip # 9 trips in one

	Time	Velocity	Elevation	Throttle	Motor Torque	Longitudinal Acceleration	Regenerative Braking Signal	Battery Voltage	Battery Current	Battery Temperature	...	Temperature Footweel Co-Driver	Temperature Feetvent Co-Driver	Temperature Feetvent Driver	Temperature Head Co-Driver	Temperature Head Driver	Temperature Vent right	Temperature Vent central right	Temperature Vent central left	Temperature Vent right.1	tripId
0	0.0	0.0	437.0	0.0	0.0	-0.34	0.0	389.9	-1.60	13.0	...	9.91	12.44	10.17	11.31	11.40	12.01	11.13	11.48	10.87	1
10	1.0	0.0	437.0	0.0	0.0	-0.30	0.0	389.9	-1.54	13.0	...	9.88	12.44	10.21	11.27	11.52	12.01	11.17	11.48	10.83	1
20	2.0	0.0	437.0	0.0	0.0	-0.34	0.0	389.9	-1.54	13.0	...	9.83	12.48	10.30	11.22	11.74	11.97	11.22	11.48	10.75	1
30	3.0	0.0	437.0	0.0	0.0	-0.31	0.0	389.9	-1.44	13.0	...	9.79	12.56	10.31	11.22	11.79	11.92	11.19	11.48	10.66	1
40	4.0	0.0	437.0	0.0	0.0	-0.33	0.0	389.9	-1.46	13.0	...	9.70	12.65	10.22	11.26	11.71	11.92	11.13	11.48	10.65	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
56060	5606.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	25.89	39.33	37.93	21.70	22.66	32.87	32.43	33.04	30.95	8
56070	5607.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	25.89	39.40	38.02	21.70	22.74	32.87	32.43	32.96	30.95	8
56080	5608.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	25.42	39.10	37.67	21.78	22.66	32.56	32.35	32.80	30.95	8
56090	5609.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	24.58	36.57	35.35	21.70	22.48	32.28	32.34	32.62	30.86	8
56100	5610.0	0.0	654.0	0.0	0.0	-0.21	0.0	389.2	92.90	14.0	...	24.49	36.27	32.99	21.70	22.40	32.26	32.34	32.61	30.86	8

19335 rows × 49 columns

trip # a single trip

	Time [s]	Velocity [km/h]	Elevation [m]	Throttle [%]	Motor Torque [Nm]	Acceleration [m/s^2]	Regenerative Braking Signal	Battery Voltage [V]	Battery Current [A]	Battery Temp	...	Temp Footweel Driver	Temp Footweel Co-Driver	Temp Feetvent Co-Driver	Temp Feetvent Driver	Temp Head Co-Driver	Temp Head Driver	Temp Vent right	Temp Vent central right	Temp Vent central left	Temp Vent right
0	0.0	0.00000	449.00000	0.0	0.00000	-0.32930	0.0	388.80000	-19.06128	7.0	...	9.30159	8.42878	7.73053	8.07965	12.53101	13.49111	10.26169	9.65072	9.65072	10.34897
10	1.0	0.00000	449.00000	0.0	0.00000	-0.42098	0.0	388.80000	-18.21941	7.0	...	9.30159	8.49573	7.99832	8.37524	12.53101	13.49111	10.26169	9.65072	9.65072	10.34897
20	2.0	0.00000	449.00000	0.0	0.00000	-0.35800	0.0	388.90000	-17.74028	7.0	...	9.30159	8.71690	8.41439	8.79826	12.53101	13.34336	10.26169	9.71767	9.65072	10.34897
30	3.0	0.00000	449.00000	0.0	0.00000	-0.34662	0.0	388.95974	-17.27922	7.0	...	9.30159	8.91179	8.91773	9.53025	12.53101	13.31655	10.26169	9.67106	9.58376	10.27508
40	4.0	0.00000	449.00000	0.0	0.00000	-0.35230	0.0	388.95976	-17.07927	7.0	...	9.37547	9.22024	9.64223	10.38966	12.45714	13.31655	10.26169	9.65072	9.56344	10.26169
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
32470	3247.0	0.57881	476.00966	0.0	4.35177	0.21247	0.0	373.54784	-3.63922	15.0	...	24.13947	24.86533	38.10452	36.27161	22.51915	22.15101	29.78516	27.63073	29.11452	28.22271
32480	3248.0	0.09375	476.00000	0.0	0.00000	-0.28795	0.0	373.60000	-3.58716	15.0	...	24.13947	24.83772	38.10452	36.27161	22.48112	22.13199	29.81277	27.69040	29.11452	28.17342
32490	3249.0	0.00000	476.00000	0.0	0.00000	-0.34132	0.0	373.60000	-3.86484	15.0	...	24.07120	24.83772	38.10452	36.27161	22.41285	22.13199	29.81277	27.71801	29.05483	28.22270
32500	3250.0	0.00000	476.00000	0.0	0.00000	-0.36000	0.0	373.60000	-3.20841	15.0	...	24.05219	24.83772	38.10452	36.27161	22.39384	22.13199	29.75311	27.71801	29.08692	28.17342
32510	3251.0	0.00000	476.00000	0.0	0.00000	-0.34236	0.0	373.60000	-3.23119	15.0	...	23.98393	24.89738	38.04486	36.20333	22.32558	22.13199	29.72549	27.71801	29.11452	28.22270

3252 rows × 48 columns

# Predictions
bTr = bigTrip.dropna() # So that we dont have to deal with missing values

from matplotlib.pyplot import figure

figure(figsize=(8, 6), dpi=80)

<Figure size 640x480 with 0 Axes>

<Figure size 640x480 with 0 Axes>

def TimePlot():
    for ind, column in enumerate(bTr.columns):
        X = bTr
        y = bTr[column]

        model = LinearRegression()
        model.fit(X, y)

        y_pred = pd.Series(model.predict(X), index=X.index)

        plot_params = dict(
            color="0.75",
            style=".-",
            markeredgecolor="0.25",
            markerfacecolor="0.25",
            legend=False,
        )

        #ax = y.plot(**plot_params)
        #ax = y_pred.plot()
        fig, ax = plt.subplots()
        plt.xlim([0,500])
        ax.plot(X["Time"], y, '.', color='0.85')
        ax.plot(X["Time"], y_pred)
        ax.set_aspect('equal')
        ax.set_ylabel('Velocity')
        ax.set_xlabel('Time')
        ax.set_title('Time Plot with Prediction');
        plt.figure()

plt.style.use("seaborn-whitegrid")
plt.rc(
    "figure",
    autolayout=True,
    figsize=(11, 4),
    titlesize=18,
    titleweight='bold',
)
plt.rc(
    "axes",
    labelweight="bold",
    labelsize="large",
    titleweight="bold",
    titlesize=16,
    titlepad=10,
)
%config InlineBackend.figure_format = 'retina'

fig, ax = plt.subplots()
plt.xlim([0, 500])
ax.plot('Time', 'Velocity', data=bigTrip, color='0.75')
ax = sns.regplot(x='Time', y='Velocity', data=bigTrip, ci=None, scatter_kws=dict(color='0.25'))
ax.set_title('Time Plot of Car Velocity');

TimePlot()

C:\Users\Chris\AppData\Local\Temp/ipykernel_7288/3072293575.py:21: RuntimeWarning: More than 20 figures have been opened. Figures created through the pyplot interface (`matplotlib.pyplot.figure`) are retained until explicitly closed and may consume too much memory. (To control this warning, see the rcParam `figure.max_open_warning`).
  fig, ax = plt.subplots()

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

<Figure size 792x288 with 0 Axes>

clear()

Tha'ts it

Doing prediction with this datasets, is near to impossible.

Why is this a bad dataset for prediction ?

First of all, we need something to predict. Predicting the velocity of
a car at a specific time is not something that is good to be predict.

Why not ?

When in traffic you can not decide your speed all the time. Sometimes there
are those pesky other drivers on the road. In really rare situtations they
even cause a so called traffic jam. We could predict when traffic jams occur,
but this would be a whole other story with a completly diffrent dataset.

What would be a good dataset ?

• Covid Cases per Day
• Birth Rates
• Cat or Dog I

Kids Project

• Projecter Owner : Chris B.
• Data Researchers
  • Chris B
  • Emil K
  • Sebastian E

Want a pdf from this Notebook ?

display(HTML('''<button id=\"print\" onclick=\"print()\" style=\"width:100%; background-color: #4CAF50; /* Green */ border: none; color: white; padding: 15px 32px; text-align: center; text-decoration: none; display: inline-block;font-size: 16px;\">Print PDF</button><style type=\"text/css\" media=\"print\">@page { size: landscape; }'''))

Print PDF